.png)
Cleaning DataVolley Scout Files From Other Analysts
tl:dr
It’s quite common to exchange DataVolley scout files (.dvw) files with other analysts on professional teams in order to prepare for them in an upcoming match
Before you start analysing it, you will need to clean the data as it will have consistency issues, such as:
You and the other analyst may use different operational definitions for the same code, e.g. your definition of a perfect pass (R#) may be different to the other analyst
You and the other analyst may use different attack combination codes. DataVolley does have a function to convert them but the other analyst may have different attack combinations that you don’t use, e.g. I don’t have a code for a one-footed front quick middle attack.
You and the other analyst may have slightly different team and player codes. This results in not being able to combine your DataVolley scout files with the other analyst’s scout files to get a combined dataset
You can clean the data to fix the consistency issues. However, I found it was quicker and cleaner to stat the match myself from the associated match video. Or I would leave it as it is if we were going to face that team once, as the data is good enough and watching the match video(s) tends to be more valuable than reviewing stats
Preface
I had this topic in my drafts for a couple of years now but I never really started it. When I wrote about cleaning data from other volleyball scouts, I was going to write everything I wanted to say on the subject. However, I soon realised it was a much bigger topic that deserved its own blog post.
Swapping DataVolley Scout Files
In higher levels of volleyball, it’s quite common to exchange DataVolley scout files (.dvw) files with other analysts (also known as scouts) on professional teams in order to prepare for them in an upcoming match. I’ve done this at European competitions, such as the Champions League, CEV Cup, and NEZVA Club Championships. I’ve also done this at the Super League level.
I found I could not simply import a swapped DataVolley scout file into my DataVolley and start analysing. This is because how I code the match may not be the same as how the other Analyst would code a match. This would lead to consistency issues, which means the data coming out of it may be what you think and you can’t make any decisions based on faulty data. So what are some of the consistency issues I encountered?
Consistency Issues
Different Operational Definitions for the Same Code If the other analyst defines an action differently to you then the data coming out of it will not be the same as how you or your team perceive it. One example is how I use the Free Ball (One example is how I use the Free Ball (F) code to mean free ball attempts whereas most pro scouts use it to mean free ball dig attempts (see ) code to mean free ball attempts whereas most pro scouts use it to mean free ball dig attempts (see link to read more). This means if I used the other analyst’s DataVolley scout file without cleaning it, the free ball stats would be completely meaningless. In addition, my to read more). This means if I used the other analyst’s DataVolley scout file without cleaning it, the free ball stats would be completely meaningless. In addition, my aggression % calculation would not work, and it would affect the point-scoring stats as I include free ball kills ( would not work, and it would affect the point-scoring stats as I include free ball kills (F#) and free ball errors () and free ball errors (F=) in the calculation. ) in the calculation.
Another example I ran into (the first one in fact) was when I used the other analyst’s DataVolley scout straight away. I found that the point-scoring stats were extremely bad. It was because the total errors committed were unusually high and the cause was due to a very high block error count. When I looked into the scout file itself, I found the scout would code both block errors (e.g. net touch, blocking the setter) and tooled blocks (attacked ball goes off the blocker’s hands for a point lost) together under the B= code. However, I would code them separately into B= and B/, respectively. This is because I don’t consider tooled blocks as true errors as I consider the opposition attacker as winning the point rather than the blocker losing the point. This was why the total block errors were so high and it adversely affected the point-scoring stats.
One last example is how does the other scout grade the pass? What is their definition of a perfect pass (R#)? If it is different to your definition then any passing stats coming out of it (pre-cleaned) will not mean the same to you or to your team. For example, if the other analyst is quite generous with their definition of a perfect pass, their passing stats would bloated compared to how you define it.
Different Attack Combination Codes Attack combination codes are two alphanumeric codes that enable specific attack types to be scouted quickly and more accurately. I use my own attack combination codes (see link), which will be different to what other scouts will use. DataVolley does allow you to convert them into the ones I use so this process is easier. This can get tricky if the attack combination name is in a different language (see screenshot below). You may also find that other scouts have different attack combination codes that I don’t have at all. For example, former USA Women’s Scout Joe Trinsey has attack combination codes for one-footed front quick middle attacks and one-footed back quick middle attacks, something I don’t code for. former USA Women’s Scout Joe Trinsey has attack combination codes for one-footed front quick middle attacks and one-footed back quick middle attacks, something I don’t code for.
Different Team Codes The more data you have on a player and team, the more reliable that data is. So I like to compile data from several matches so I get a better idea of the team or player’s real performance level. However, if your team code is different to the other scout’s team code for the same team, then you are unable to combine your and the other scout’s stats together. This happened to me when I used a ‘TN1’ team code for Team Northumbria Women, which was used by Volleyball England, but their scout used a ‘TNW’ team code to distinguish it from the Men’s ‘TNM’ team code.
Different Player CodesThe team code issue applies to player codes. If the same player has two different player codes, one from me and one from the other scout, then their name and respective stats will appear twice (assuming the team code is the same). This has happened to me due to a spelling mistake. The other scout had the player code like BLO-JOE whereas I had it like BOL-JOE.
Your Bespoke Worksheets and Reports May Not WorkAll of these differences may mean some of your worksheets and reports may not work. I already mentioned that any free ball stats shown will be completely meaningless. I also talked about how the higher block error count (since it includes block-outs) will adversely affect the point-scoring stats that it would be meaningless.
A different example is I used another analyst’s DataVolley scout file and the setting distribution zone chart was blank. After some investigation, it was because they didn’t code attack directions and the zone chart relies on the attack origin to get the setting distribution.
I also have a player report, that shows attack directions by zone (left, right, pipe, 10 for wing players; quick front, back quick, shoot, slide for middles), serve direction plus some other stats. If the other analyst uses different attack combination codes to me then those attack directions charts will be blank because it is specific to my attack combination codes. Also, if serve directions are not coded then the serve direction chart will be blank.
Cleaning the DataVolley Scout File
You can clean the DataVolley scout file prior to analyse the data on it. There are several things you’ll need to do.
Convert Attack Combination CodesYou would do this when you open up the DataVolley scout file. You assign their attack combination code to your own. If they use an attack combination code that you don’t use then there’s not much you can do about it
Search for Codes to Manually Edit it and Insert New CodesYou use the search function to manually edit it. For example, I would look for all B= codes and change it to a B/ if it was a block out. I would also use the search function to insert new codes. For example, I would search for all F (Free Ball) codes to change it to an emergency dig (DE) code (my code for free ball digs) and I would insert an F code for a free ball attempt.
Change the Team to Your Version of the TeamYou can change the team in the match notes from the other scout’s version of Team 1 to your version of Team 1. This is so you won’t get conflicts with the team and player codes. I would do it for both teams in the match, especially if you are going to face them in the future. You do have to make sure the shirt numbers are correct so that what was coded by the other analyst corresponds to the correct player. You also have to input the starting rotations again plus substitutions (*) in the players list so the rotation stats work. It may be best to take a screenshot of the player list before changing over.
Use the Standard DataVolley ReportsSince my bespoke worksheets and reports may not work properly, using the standard DataVolley match reports should work with any scout file, without too many errors. They use standard calculations so players and coaches should still understand it.
In Practice
I’m not going to lie but cleaning other analysts’ DataVolley scout files took way too long. It would have been quicker if I just coded the match myself from the associated match video. I also know it would be much cleaner to stat it myself (first-party/ internal data) and I wouldn’t encounter any issues down the line. This is if I know we will face the team(s) again later in the season because I can use the data again. However, if we were only going to face a team once, e.g. European matches, then I tend to leave the DataVolley scout file as it is. The data is good enough and watching the match video tends to be more valuable than reviewing the stats.
Comments